Arsenic is of concern because it is a Superfund toxicant and a human carcinogen. It is one of the few human carcinogens which is not tumorigenic in rodents. Human cells are much more sensitive to the cytotoxic effects of arsenite (the likely carcinogenic form) compared to rodent cells, and there is evidence for heterogeneity in the response of humans to arsenic compounds. The main hypothesis to be tested is that cellular efflux of arsenic ("arsenic pump") is a primary mechanism of arsenic resistance, and heterogeneity in cellular efflux genes may underlie differences in susceptibility to arsenic compounds. Thus, the goals of this proposal are a molecular description of the pathway responsible for arsenite efflux in human cells and a determination of genetic heterogeneity in the genes encoding the proteins making up the pathways. We have previously shown that mammalian cells do have at least one arsenite efflux system. The product of a yeast (S. cerevisiae) gene, ACR3, is most likely an arsenite efflux pump with a function similar to the bacterial ArsB proteins. We have also demonstrated that the S. cerevisiae ACR2 gene is required for arsenate resistance and encodes an arsenate reductase which reduces arsenate to arsenite (substrate of the pump). No human homologues in human cells, we will use alternative methods which we have used to clone other resistance to arsenite (substrate of the pump). No human homologues to these sequences have been identified to date. A human liver cDNA library will be used to PCR amplify sequences with degenerate ACR3 primers. In the event that there are no ACR3 homologues in human cells, we will use alternative methods which we have to clone other resistance genes in this laboratory: expression cloning and/or PCR-select differential screening. We have isolated arsenite- resistance Chinese hamster and human osteosarcoma cells for use in these alternative strategies. We will also perform expression cloning by transfecting human cDNAs into yeast strains which have disrupted ACR2 and ACR3 genes. Any cDNAs isolated by any strategy will be transfected into wild type human cells in both sense and anti-sense orientation, and the resulting phenotypes will be characterized. Finally, we will compare arsenite efflux rates and genotypes in human cells from different donors which have different susceptibilities to arsenic compounds.